This project aims to address a problem faced by many new medical students. When students first begin to learn clinical medicine, they learn about one or two diseases at a time, collecting knowledge aboutthe symptoms, signs and investigations associated with each disease. For example, they are taught that myocardial infarction is associated with chest pain – and when they see their next patient with chest pain they inevitably make the diagnosis of a heart attack. It is only as they learn about more and more diseases and begin to gain an appreciation of the relative importance of different symptoms and signs associated with each condition that they begin to learn the complexities of medical diagnosis.
The Medical Knowledge Base is a project developed by Mohammad Ubaydli, a clinical student on the Cambridge Clinical Course, in collaboration with the Clinical and Biomedical Computing Unit. The aim has been to develop a program that can help students and junior doctors make the associations between symptoms and signs, investigations and diagnoses. There are several different components.
The database is used to record the details of an individual disease in a structured way into forms with fields for symptoms and signs, investigations, epidemiology, treatment etc. Each symptom, or sign can be weighted as to whether it occurs nearly always, frequently, sometimes or rarely in a given condition. Freetext notes can be used to qualify any piece of information entered into the fields. Each disease is entered in isolation of any other in the database.
The ‘wobulator’ view displays the data in the database in a graphical format and shows the relationships of each disease to its symptoms, signs and investigations. Each feature in this display can be used as a ‘pivot’; thus for the entire database for example, if the key disease is cyanotic heart diseases, one of the signs would be clubbing, simply clicking on clubbing rotates all the data from the database and shows all the other diseases that are associated with clubbing. None of these links and associations need to be done manually because the entire process is automated by the back-end database.
The focuser allows the displayed data to be refined so that the user can hide the less-common associations or symptoms and signs from the display. For example, if shortness of breath is a rare symptom of a hiatus hernia, then the focuser can be used to exclude it from the display of conditions associated with shortness of breath so that only those conditions where shortness of breath is very common are displayed.
The notes extractor:
This allows a student to extract the data about individual diseases from the database into an HTML file, so that they can be accessed as notes through a webbrowser.
The development of a collaborative learning environment:
The Medical Knowledge Base encourages collaborative working among groups of students, because each student takes responsibility for individual diseases and enters the data in the database. The Medical Knowledge Base automatically creates the links to the information inserted by other students; the holistic environment this creates becomes more valuable for the group as a whole.
- Team leader: Jem Rashbass
- Technologies: Java
- Funding: Core funding from CBCU
- Start: September 1997